Author Topic: Greg's Bandit 400 Fuel Injection Project...  (Read 146228 times)

Offline bdouvill

  • Everybody
  • Board Homesteader!
  • *
  • Posts: 78
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #165 on: December 10, 2014, 11:39:55 AM »
Quote
Got it even without the picture ;-)
Funny to notice the whole packaging is the same, only the switch is not present and there's a cover on it. I believe that symplifies the number of references for the manufacturer.

I really wasn't trying to be pedantic (or to offend), just documenting the headlight issue now that I've modified the system to a configuration that's pleasing to me. 


You were not at all, "au contraire" ;-)
Picture is better than any explanation for me. Just funny to see how Suzuki handled that. And I was not expecting a motorcycle from the early 90's to get such a feature.

I hope you can ride it soon and enjoy it. Let us know ;-)
Benoit.
Bandit 400 1992, 1st bike.

Offline greg737

  • Board Homesteader!
  • ***
  • Posts: 315
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #166 on: December 10, 2014, 11:14:23 PM »
It rained all day long yesterday but this morning (Wednesday) it stopped and the sky cleared. 

I waited all day for the streets to dry.  While waiting I got the last bits of idle-speed and cold start-up tuning completed, now the Bandit starts and idles beautifully.  Then, as mid-afternoon arrived, I reassembled the bike to make it rideable for the first time since October of 2013. 

I put the airbox, fuel tank and rider-seat back on the bike, added a couple of gallons to the fuel tank, corrected the tire pressures and put on a tail-bag to carry the laptop computer.

The computer would be hooked up to the Microsquirt during the ride, running a piece of tuning software called TunerStudio so I could data-log the ride for review afterwards.

Don't know how you guys react to this sort of moment but for me it's alway a very nervous time when I'm out on a first-time road test: Will it run great?  Or will it burn to the ground two miles from home?  Success or Failure?

I knew ahead of time that the main fueling table (the Volumetric Efficiency Table, a.k.a. "the V.E. table") was only a rough approximation of what I thought the bike would need so the first ride was bound to have some odd hesitations and jumpy, lumpy feeling throttle responses. 

The first mile went okay, but with some weird fueling as expected so I pulled over and turned on the TunerStudio feature called "Tune Analyze Live" which automatically corrects fueling as you ride.  This feature references (moment by moment, in real time) the Air/Fuel ratio table to learn what the correct ratio should be and compares that number with what the wideband oxygen sensor is currently reporting.  If the two differ the program makes incremental corrections to eventually perfect the V.E. table.

And now, here's my screwup of the day:
When I rode out of my neighborhood and onto a nice straight road I began to slowly open the Bandit up, cautiously taking the RPMs higher.  At 6,000 RPM the bike stumbled, as if the ignition had momentarily quit.  The engine didn't fail, but it wouldn't rev past 6,000 RPM.  It was a bit worrying so I cut this first test-ride short and took the Bandit back home.

When I got to the garage and looked at the TunerStudio data-log I discovered the rev-limiter parameter in the software was set to 6,000 RPM.  I didn't set it there, this was a default-setting that I never got around to changing.  So there was nothing wrong with the Bandit.

Now I'll have to wait out the next two days of rain for another chance at a test-ride.  I plan to shoot a bit of video on that ride.
« Last Edit: December 10, 2014, 11:16:17 PM by greg737 »

Offline andrewsw

  • Board Homesteader!
  • ***
  • Posts: 598
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #167 on: December 11, 2014, 12:02:27 AM »
 :clap:

man, I'm so anxious for the next installment. Thanks for sharing all this.

A

Offline greg737

  • Board Homesteader!
  • ***
  • Posts: 315
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #168 on: December 11, 2014, 08:40:42 PM »
I thought I would try taking a screen-shot from the TunerStudio data-log that I captured during the Bandit's first test ride.  I'm not sure how legible it will be...

This is a screen-shot of the real-time data traces.  It captures the moment when I first ran into the 6,000 RPM rev limit that I'd failed to change (the TunerStudio software comes from the manufacturer with the REV LIMIT parameter set to 6,000 and you have to change it to match your engine, but I forgot to do that).
.

.
You can see that the data-log is made up of three parallel boxes that move through time from left-to-right, each box has traces of three different engine parameters in different colors. 

In the top box the WHITE line is RPMs and you can see that when the engine hits 6,000 everything goes haywire. 

In the middle box the RED trace is AIR/FUEL RATIO (as sensed by the Bosch wideband oxygen sensor on the Bandit's exhaust header).  When the engine hits the rev limit parameter of 6,000 the Microsquirt ECU shuts off fueling by reducing the commanded PULSE WIDTH (how long the injectors are held open) to nearly zero, which is shown in the middle box by the sudden decrease in the GREEN trace.  As a result the RED trace (AIR/FUEL RATIO) jumps upward and for a moment the engine feels like it's going to quit.

As soon as this happened I got off the throttle, which is shown by the GREEN trace in the top box suddenly going downward (TPS = THROTTLE POSITION SENSOR).

So the RPMs drop below 6,000 the Microsquirt resumes commanding normal fueling and everything goes back to normal.

This is what it looks like when you bounce an engine off a rev limiter

(to be specific, this is what a simple, old-school-style fuel-limiting rev limiter looks like.  Modern OEM ECUs are much more gentle and progressive when they invoke rev limits, usually using a combination of ignition retard and partial fuel cut).
« Last Edit: December 11, 2014, 08:53:59 PM by greg737 »

Offline Squishy

  • Board Homesteader!
  • ***
  • Posts: 417
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #169 on: December 12, 2014, 11:01:28 AM »
 :clap:
Very interesting.

Keen to see the results and video with proper rev limiter

Offline greg737

  • Board Homesteader!
  • ***
  • Posts: 315
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #170 on: December 12, 2014, 05:52:00 PM »
:clap:
Very interesting.

Keen to see the results and video with proper rev limiter

Thanks.  As soon as it stops raining and the streets dry I'll be back out for more test-rides.

Road testing is definitely the most enjoyable, exciting part of this project.  There's nothing like twisting the throttle, is there?  But today, which is very rainy, I'm stuck in the garage doing what amounts to tuning-drudgery, the boring stuff.

So, how about a series of short videos of a boring cold startup test?  (I've got a lot of space on my Photobucket account so I'm not being shy about shooting and posting video.  As I've mentioned before, I'm just the sort of techno-gearhead-nerd who finds all of this very interesting).

.
Video: http://s679.photobucket.com/user/EWflyer/media/GSF400%20vid%2011_zps5qrankya.mp4.html?sort=3&o=2
.
Video: http://s679.photobucket.com/user/EWflyer/media/GSF400%20vid%2012_zpshpivybwa.mp4.html?sort=3&o=1
.
Video: http://s679.photobucket.com/user/EWflyer/media/GSF400%20vid%2013_zpsuk5relcj.mp4.html?sort=3&o=0

Offline Squishy

  • Board Homesteader!
  • ***
  • Posts: 417
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #171 on: December 13, 2014, 11:35:05 AM »
The more video's the better, in my opinion  :thumb:

Just some things I noticed from video's: (you probably know this but still).
- 12.00V is pretty low for a battery.
- What makes you say it needs more air at start? The more choke I use at start, the higher the RPM gets to initially. It shoots right up to 2000rpm. Without choke, it starts like how it was on your video.
- Are you running without exhaust silencer? It sounds like at that at low rpm or like you have a leak on the exhaust.
- Why was it increasing RPM because the coolant was hot? Doesn't make sense to me.

Offline greg737

  • Board Homesteader!
  • ***
  • Posts: 315
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #172 on: December 13, 2014, 03:30:47 PM »
The more video's the better, in my opinion  :thumb:

Thanks, I appreciate your vote in favor of more videos (and photos).

Quote
Just some things I noticed from video's: (you probably know this but still).
- 12.00V is pretty low for a battery.

Yes, you're right, the battery is probably nearing the end of its useful lifespan.  The battery came with the bike when I bought it and like the rest of the bike the battery's history is unknown to me.  I'm trying my best to preserve this battery well enough to get through the building and testing phase of the project, I don't know if you noticed but I've been using a large farm-tractor battery and jumper cables to supplement the Bandit's battery for stationary testing.  Doing this has allowed me to do the type of testing that requires multiple extended cranking/starting-stopping of the Bandit's engine without flatting the Bandit's battery.  When the work is all done and the project is 99% complete I'll treat the Bandit 400 to a nice new battery.



Quote
- What makes you say it needs more air at start? The more choke I use at start, the higher the RPM gets to initially. It shoots right up to 2000rpm. Without choke, it starts like how it was on your video.

It sounds like you might not be totally familiar with the layout of the Mikuni BST33 slingshot carburetor, especially its Choke circuit.  The BST33 carb mixes BOTH air and fuel in its choke circuit.  Let me walk you through the choke circuit of a stock, original Mikuni BST32 (which is exactly the same as the BST33 except for the size of its bore)...
.

.

.

.

.

.

.

.

.

.

.

.

.
.
.
Quote
- Are you running without exhaust silencer? It sounds like at that at low rpm or like you have a leak on the exhaust.

My Bandit's exhaust system is an aftermarket Yoshimura "4 into 2 into 1" style header and the exhaust canister is also a Yoshimura.  The canister is free-flowing, it doesn't have an add-on silencer.  The Yoshimura header is of the type designed to maximize scavenging/flow at mid-range power/torque, in its first two "Y"s it combines cylinders #1 and #4 together and cylinders #2 and #3 together.  Then it combines these two at its second "Y".

I had the header sand-blasted and ceramic coated so it looks nice but the canister is still in its original (to me) state which is very dented and scraped up.  There are no leaks in the system.  If the exhaust note sounds odd on the video I'd guess it's because the sound is bounding around in my small garage.


Quote
- Why was it increasing RPM because the coolant was hot? Doesn't make sense to me.

Okay, maybe my word choice on the video was a bit confusing.  When I say, "because the coolant temperature is rising" on the video I'm really thinking, "because the temperature of the metal, plastic and rubber in the Bandit's intake tract is rising".  (Obviously there are other parameters to consider; for example, the temperature of the oil affects how it flows.  Cold oil is a drag on the system that has to be compensated for.)

Starting an engine from dead cold is a complex, dynamic situation. 

When you start up from dead cold the metal and any other substances, like plastic or rubber used in the walls of the intake tract (carbs, connector boots and the cylinder head) are cold and the intake fuel vapor readily adheres to them.  Because of this the cold startup cycle (from dead-cold to warmed up to normal operating temperature) is a highly dynamic, constantly changing set of parameters.

In modern fuel injection systems the engineers who design them actually take into account the amount of fuel that can/will adhere to the metal in the intake tract.  They take everything into account: the temperature, the amount of metal, the metal's thermal characteristics, the amount of fuel being injected, the volume of air being moved.  And they come up with a formula that contains variables that account for both "fuel that adheres to walls" and "fuel that is sucked from walls".

This necessitates a very carefully choreographed fueling scheme to achieve smooth RPM control during a fuel injected startup.  When the enormous computational ability of a human brain is placed in charge of a Choke lever during cold startup this complexity is easily handled, almost as an afterthought.  But when you're programming a computer to do it you have to be very elementary about it.  Coolant temperature is the parameter we use to estimate the heat-state of the intake tract. 

Obviously, as the temperature rises these two variables ("adheres to walls" and "sucked from walls") are in a constant state of change which means the air/fuel ratio is also going to constantly change.  In the early part of the start, when things are at their coldest the "adheres to walls" factor is the biggest, but as the temperature of the intake tract rises toward normal operating temperature the "sucked from walls" factor becomes the biggest factor. 

When normal sustained operating temperature is reached the two reach and maintain an almost balanced state, except during rapid throttle changes.
« Last Edit: December 13, 2014, 03:40:49 PM by greg737 »

Offline vintagemilano

  • TURBO TYPER!
  • **
  • Posts: 21
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #173 on: December 13, 2014, 11:00:15 PM »
Here is another vote for pictures/video/updates. Love it all.
1993 GSF400
Cranbrook, BC, Canada

Offline greg737

  • Board Homesteader!
  • ***
  • Posts: 315
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #174 on: December 14, 2014, 10:09:14 PM »
I had a pretty tough day in the garage. 

Over the last 3 days I've begun taking the Bandit out for test rides.  It wasn't going well at all.  The engine wouldn't rev past 6,000 RPM.

After the first test ride I thought it was a mistake I'd made with setting the rev limiter in TunerStudio, but after I fixed the mistake the problem still remained.

At 6,000 RPM the Microsquirt ECU was losing "Sync".  Losing Sync means the stream of inputs that the Microsquirt ECU was getting from the pulse generator (the output from the VR sensor and trigger wheel) were not matching with the ignition trigger definition parameters that I had entered.  If the Microsquirt cannot continuously verify the engine's crankshaft position it "loses sync" and the engine feels like you're hitting a rev limiter.

Today I went to the MSExtra website forum and asked for help.  I was pretty lost but there are some very knowledgable guys over there and they looked at my datalogs and gave me some very helpful thoughts and advice.

If you're interested in having a look at the neediness and insecurity that lurks behind my normal facade of intelligence and capability, here's a link to the forum thread that I started this morning: http://www.msextra.com/forums/viewtopic.php?f=101&t=56874.

In just a page and a half of forum posting I went from hopelessly lost to problem solved!

Now the Bandit revs like it should.  I hope to have a video of it soon.

Offline vintagemilano

  • TURBO TYPER!
  • **
  • Posts: 21
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #175 on: December 14, 2014, 11:34:19 PM »
K, that's some pretty cool stuff. Nice work! 
1993 GSF400
Cranbrook, BC, Canada

Offline greg737

  • Board Homesteader!
  • ***
  • Posts: 315
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #176 on: December 15, 2014, 01:12:59 AM »
Here's a very short video I'm calling "The Sound of Failure"...

Video:http://s679.photobucket.com/user/EWflyer/media/GSF400%20vid%2014_zpsfmlvosyz.mp4.html?sort=3&o=0

This is a short bit of video from the second day of test riding.  At this point the Bandit still wouldn't rev beyond about 6,000 RPM.  I was pretty down about the situation at the time, but now that I've solved the problem I'm feeling much better about how bad the Bandit was running in this clip.  And it was bad, the bike even quits completely at the end.  I had to coast it into a U-Store-It parking lot.
« Last Edit: December 15, 2014, 01:15:23 AM by greg737 »

Offline greg737

  • Board Homesteader!
  • ***
  • Posts: 315
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #177 on: December 15, 2014, 05:27:36 PM »
More video?  Yes... more video.

Today's theme is "The sound of something approaching or at least moving toward success"

Just a couple of video clips from today's ride. 

Now that the Microsquirt is getting perfect signals from the pulse generator (VR sensor and trigger wheel) the Bandit is running great.


Video: http://s679.photobucket.com/user/EWflyer/media/GSF400%20vid%2015_zpsrke4xtk6.mp4.html?sort=3&o=1

Video: http://s679.photobucket.com/user/EWflyer/media/GSF400%20vid%2016_zpsq4uwzbsv.mp4.html?sort=3&o=0

I'm riding the Bandit carefully because the fueling table is still rough and there are spots where it over-fuels and spots where it under-fuels.  During this ride I had the laptop computer in the Bandit's tailbag.  The laptop was running TunerStudio and I had the TuneAnalyze Live function turned on.  The TuneAnalyze Live function actually tunes the fueling table while you ride the motorcycle.
« Last Edit: December 15, 2014, 05:32:12 PM by greg737 »

Offline Squishy

  • Board Homesteader!
  • ***
  • Posts: 417
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #178 on: December 15, 2014, 06:08:58 PM »
There was something else wrong with the pulse signal? (apart from it being too wide?)

Still sounds like you don't have an exhaust muffler  :yikes:

Did you go full throttle in the 2nd video?

Offline andrewsw

  • Board Homesteader!
  • ***
  • Posts: 598
Re: Greg's Bandit 400 Fuel Injection Project...
« Reply #179 on: December 15, 2014, 06:19:47 PM »
 :clap: :clap: :clap: :clap: :clap: :clap: :clap: :clap: :clap: